Configuring the Speech Server engine

server.manager.gracefulremoval

Specifies how the server behaves during shutdown.

Boolean

0 means immediate shutdown.

1 means graceful shutdown—the server rejects new connections and waits until active connections end.

DEFAULT: 0

server.mrcp2.sip.maxCountOfSession

Specifies the maximum number of concurrently active SIP server sessions.

Integer: 1–10,000.

DEFAULT: 96

server.threadmanager.initialcount

Specifies the initial number of worker threads started by the server.

Integer

DEFAULT: 1

server.threadmanager.maxcount

Specifies the maximum number of worker threads in the server.

Integer: 1–INT_MAX threads.

Typically, the number must be equal to the planned number of concurrently served sessions.

DEFAULT: 250

server.mrcp1.rtsp.maxCountOfSession

Specifies the maximum number of concurrently active server sessions.

Integer: 1–10,000.

DEFAULT: 500

server.mrcp1.rtsp.sessionTimeout

Specifies the session timeout for RTSP transport.

Integer: Maximum time (in milliseconds) to wait for the next message in a session.

DEFAULT: 60000 (1 minute)

server.mrcp1.transport.port

Specifies the session RTSP port to use for the server.

Integer. An available port number.

DEFAULT: 4900

server.mrcp1.transport.TCPTimeout

Specifies the maximum number of seconds to wait for activity before closing the TCP connection, after the session was deleted on session timeout.

Integer: seconds. Minimum 1s 0; there is no maximum.

DEFAULT: 180

server.mrcp2.sip.logLevel

Specifies the logging level of the SIP stack.

Integer: -1 to 8 (9 and 666 for Nuance internal use)

DEFAULT: -1 (no SIP logging)

server.mrcp2.sip.maxCountOfSession

Specifies the maximum number of concurrently active SIP server sessions.

Integer: 1–10,000.

DEFAULT: 96

server.mrcp2.sip.sessionTimeout

Specifies the maximum time to wait for the next message in a session.

Integer: 0–INT_MAX (milliseconds)

DEFAULT: 60000 (1 minute)

server.mrcp2.sip.transport.tcp.port

Specifies the SIP TCP port to use for the Speech Server.

An available port number

DEFAULT: 5060

server.mrcp2.sip.transport.udp.port

Specifies the SIP UDP port to use for the Speech Server.

An available port number

DEFAULT: 5060

server.mrcp2.transport.tcp.port

Specifies the MRCP2 port to use for the server.

An available port number

DEFAULT: 6075

server.mrcp2.transport.timeout

Specifies the maximum timeout before receiving the next data from a TCP (TLS) socket.

Integer: seconds. Minimum is 0; there is no maximum.

DEFAULT: 20 (seconds)

server.rtp.bufferSize

Specifies the audio buffer size for RTP audio reception.

Integer: 1000–INT_MAX (milliseconds)

DEFAULT: 5000

server.rtp.maxCountOfSession

Specifies the maximum number of concurrently active RTP server sessions.

Integer: 1–<maxsessions>, where maxsessions is the maximum of the following three parameters:

• server.rtp.maxCountOfSession
• server.mrcp1.rtsp.maxCountOfSession
• server.mrcp2.sip.maxCountOfSession

DEFAULT: 600

server.rtp.maxPacketSize

Specifies the maximum size of accepted RTP packets.

Bytes. Minimum is 0; there is no maximum.

DEFAULT: 1024

server.rtp.port

Specifies the starting port for RTP communication.

An available port number

DEFAULT: 7892

server.rtp.strictSdpMediaPortUse

Specifies whether the server accepts packets from ports other than the UDP port and IP address.

Boolean

1 (default): rejects packets from any other port.

0: the server does not check the UDP port, and accepts packets from any port.

DEFAULT: 1

MRCP

server.mrcp2.transport.qos

Integer: decimal DSCP value representing the quality of service desired.

DEFAULT: 96 (CS3)

SIP

server.mrcp2.sip.transport.qos

Integer: decimal DSCP value representing the quality of service desired.

DEFAULT: 96 (CS3)

RTP

server.rtp.transport.qos

Integer: decimal DSCP value representing the quality of service desired.

DEFAULT: 184 (EF)

server.toneDetector.configFile

Enables tone detection and points to the configuration file that Speech Server uses to configure the tone detection library.

Path. For example: $NSSSVRSDK/config/tonedef-en-us.xml.

DEFAULT: $NSSSVRSDK/config/tones.all.on

The parameter has no default value, but the NSS baseline configuration file sets a value at installation.

server.toneDetector.originator

Specifies the software that uses the tone detection library. The <originator> component of the event name. A typical value for this parameter is "nss".

String. For example: server.toneDetector.originator VXIString nss

DEFAULT: nss

server.toneDetector.timeout

Specifies the interval after which the tone detector stops operating.

Integer. 0–INT_MAX (milliseconds)

DEFAULT: 20000

server.nrs.serverAddress

Specifies the address of a single Nuance recognition server.

Server address of the form host:port

DEFAULT: localhost:8200

server.nvs.Address

Specifies the address of a single NuanceVocalizer host.

Host address of the form host:port

DEFAULT: localhost:9200

server.nlps.serverAddress

Specifies the address of a single Natural Language Processing service host or instance.

Location (WebSocket URI) of the Natural Language Processing service.

DEFAULT: wss://localhost:8443/nlps

nlps-audio-only

Enables Dragon Voice for Krypton-only recognition (raw recognition with no semantic interpretation).

True or False

DEFAULT: False (disabled)

Enables Dragon Voice for Krypton-only recognition (raw recognition with no semantic interpretation).

Boolean

DEFAULT: 0 (disabled)

server.rtp.wcr.enable

Enables whole call recording.

Boolean

DEFAULT: 0 (disabled)

server.rtp.wcr.maxminutes

Specifies the maximum duration of whole call recording. Whole call recording must be enabled first by setting server.rtp.wcr.enable to 1.

Integer: minutes. Minimum is 0; there is no maximum.

DEFAULT: 60 (highest recommended value)

server.rtp.wcr.outputtype

Specifies whether the whole call recording is saved as a single file or split into separate files, with each file containing audio from a different party. Whole call recording must be enabled first by setting server.rtp.wcr.enable to 1.

Integer:

• 0: Writes 1 file per mid (media identifier).
• 1: Writes 1 file per session.
• 2: Writes 1 file per speaker. For example, for a session with a single mid, one file is generated for the inbound speaker and one for the outbound speaker.

DEFAULT: 0

server.rtp.wcr.sampling

Specifies the percentage of calls that are randomly selected for recording. Whole call recording must be enabled first by setting server.rtp.wcr.enable to 1.

Integer: 0–100.

DEFAULT: 1 (percent)

server.inet.extensionRule.extension

Maps file extension to a MIME type.

MIME type that corresponds to the extension in the parameter name.

DEFAULT: (none)

server.inet.userAgent

Specifies the user agent name presented to the web server when HTTP requests are made.

String

DEFAULT: NSS-MRCP/6.1

switts.inet_timeout_download

Specifies the default timeout for downloading a URI (from open through the final byte).

Integer: 1–INT_MAX (milliseconds)

DEFAULT: 30000 (30 seconds)

switts.inet_timeout_io

Specifies the default timeout for reading or writing a block of data over a web server connection.

Integer: 1–INT_MAX (milliseconds)

DEFAULT: 30000

switts.inet_timeout_open

Specifies the default timeout for opening a connection to a web server.

Integer: 1–INT_MAX (milliseconds)

DEFAULT: 30000

Rules for using proxy servers to fetch documents from application servers using HTTP.

DEFAULT: (none)

Rules for using proxy servers to fetch documents from application servers using HTTPS.

DEFAULT: (none)

server.inet.nonProxyHosts

Specifies the list of servers to be accessed directly, rather than through a proxy server.

List of host names separated by commas.

DEFAULT: (none)

server.inet.proxyPort

Specifies a port for the HTTP proxy server.

Integer. An available port number.

DEFAULT: 1111

server.inet.proxyServer

An HTTP proxy server host used by Speech Server when fetching audio files from an application server.

Host name or IP address of the proxy server host.

DEFAULT: myHost

server.callLog.enableDiskLog

Enables call logging. When disabled, no call logs are written.

Boolean

DEFAULT: 1 (enabled)

secure_context

Parent directory where the system writes call logs.

Deprecated. Use swirec.secure_context or switts.secure_context instead.

DEFAULT: $NUANCE_DATA_DIR

server.callLog.tcp.port

Specifies the listening TCP port assigned to the call log server.

TCP port number

DEFAULT: 10101

Specifies the parent directory where the system writes call logs.

Directory path. The directory does not need to exist in advance.

DEFAULT:

$NUANCE_DATA_DIR/callLogs

or

$NUANCE_DATA_DIR/CompanyName/callLogs/ApplicationName/yyyy/mmMonth/dd/hh

server.callLog.cleanupCallLogServer

Enables removal of old log files (audio and text).

Boolean

DEFAULT: 1 (enabled)

server.callLog.cleanupBreakTimeMinutes

Specifies time for the call log server cleanup process to pause between cleaning cycles.

Amount of time (minutes) that the call log cleanup process pauses between cleaning cycles. Maximum pause is 10080 (7 days).

DEFAULT: 1 (minute)

server.callLog.cleanupRates.0

Controls how long to keep each type of log file before deletion.

A string of fields (separated by semi-colons) where each field indicates a category or type of log file followed by a colon and a time designation (the rate). Set individually for each file type.

DEFAULT: Depends on file type.

server.callLog.minFreeDiskMB

Specifies the minimum available disk space for writing call logs.

Integer in MB: 10 to MAXINT on your system. The minimum value is 10 MB. There is no upper limit.

When the available space falls below this amount, logging is disable for the session until space is freed. The space is checked approximately every 20 minutes.

DEFAULT: 1073 MB

server.callLog.warnMinFreeDiskMB

Threshold to warn operators that the available disk space for call logging is too low.

Integer (megabytes): 10 to the value of MAXINT on your system. There is no upper limit.

DEFAULT: 2146 (2 gigabytes)

server.callLog.tls.port

Specifies the listening TLS port for clients to communicate with call log server.

Integer. An available port number for TLS.

DEFAULT: 10102

server.callLog.utteranceSampling

Specifies the percentage of randomly-selected utterances saved as audio files.

Integer: 0–100.

DEFAULT: 100

swirec_diag_tags_enable

Controls the logged output by enabling tags in diagnostic log files.

Integers representing one or more tags in the TRC tagmap files.

DEFAULT: (none)

server.log.contentDir

Specifies the base directory for diagnostic logging in per-company mode.

Directory path name

DEFAULT: $NUANCE_DATA_DIR/system/diagnosticLogs

server.log.filename

Specifies the location for Speech Server and Vocalizer diagnostic files.

String

DEFAULT: $NUANCE_DATA_DIR/system/diagnosticLogs/nss.log.

server.log.contentTotalSizeMB

Specifies the maximum size to allow for the MRCP message log directory.

Integer: 0–2047 megabytes.

DEFAULT: 50 (MB)

server.log.maxLogSizeMB

Specifies the maximum size for Speech Server and Vocalizer diagnostic files.

Integer: 0–2047 megabytes.

DEFAULT: 50 (MB)

server.log.keepLogFileOpen

Keeps the log file open between writes for faster logging.

Boolean

DEFAULT: 1

server.log.logToStdout

Writes logs to standard out as well as to a file.

Boolean

DEFAULT: 1 (standard out as well as to a file)

server.log.diagTag.xxxx

Specifies a particular aspect of a component to include in the Speech Server diagnostic log (nss.log)

Boolean. 1 means to include diagnostic information for the module specified by the index xxxx.

DEFAULT: 0 (tag not logged)

server.log.errorMapFile.x

Writes diagnostic information for different components to separate log files instead of the main diagnostic log.

String. The value x specifies error maps for different components as follows:

0   $NSSSVRSDK/config/serverErrors.xml
1   $NSSSVRSDK/config/osrspeechrecogErrors.xml
2   $NSSSVRSDK/config/rsspeechsynthErrors.xml
3   $NSSSVRSDK/config/inetErrors.xml
4   $NSSSVRSDK/config/osrrecorderErrors.xml
5   $NSSSVRSDK/config/dictationErrors.xml|
6   $NSSSVRSDK/config/prsspeechrecogErrors.xml

DEFAULT: (none)

server.log.extraLogging

Enables additional logging: logs a diagnostic message on every enter and return from a function.

Boolean

DEFAULT:0 (no extra logging)

server.log.reportErrorText

Reports the error texts from error map files.

Boolean

DEFAULT: 1 (report error text)

server.log.diagLogPerCompany

Writes diagnostic information to separate log files (per company) instead of the main diagnostic log.

Boolean. 0 means a single diagnostic log covers all companies.

DEFAULT: 0 (disabled, a single log is written)

server.log.secureDiagLogOSRcontext

Specifies whether recognition messages from the browser appear in Speech Server diagnostic logs.

Boolean

DEFAULT: 0 (disabled, browser messages written)

server.log.secureDiagLogTTScontext

Specifies whether text-to-speech messages from the browser appear in Speech Server diagnostic logs.

Boolean

DEFAULT: 0 (disabled, browser messages written)

server.log.suppressSensitiveDiagLogs

Specifies whether confidential data is suppressed in Speech Server diagnostic logs.

Boolean

DEFAULT: 1 (enabled, secure data suppressed)

server.log.suppressSensitiveURIs

Specifies whether Speech Server suppresses URI strings that contain confidential data.

Boolean

DEFAULT: 0 (disabled, URIs are written)

server.log.secureDiagLogOSRcontext

Specifies whether recognition messages from the browser appear in Speech Server diagnostic logs.

Boolean

DEFAULT: 0 (disabled, browser messages written)

server.log.secureDiagLogTTScontext

Specifies whether text-to-speech messages from the browser appear in Speech Server diagnostic logs.

Boolean

DEFAULT: 0 (disabled, browser messages written)

server.log.suppressSensitiveDiagLogs

Specifies whether confidential data is suppressed in Speech Server diagnostic logs.

Boolean

DEFAULT: 1 (enabled, secure data suppressed)

server.log.suppressSensitiveURIs

Specifies whether Speech Server suppresses URI strings that contain confidential data.

Boolean

DEFAULT: 0 (disabled, URIs are written)

swirec.mute_wcr

Replaces confidential data in whole-call recordings with silence.

Boolean

DEFAULT: (none)

swirec.secure_context

Sets recognizer security levels for protecting confidential data, typically for a single event.

open, encrypt, suppress

DEFAULT: open (no security)

switts.mute_wcr

Replaces confidential data in whole-call recordings with silence.

Boolean

DEFAULT: (none)

switts.secure_context

Sets security levels for protecting confidential data in logs of text-to-speech conversions, typically for a single event.

open, encrypt, suppress

DEFAULT: open (no security)